Results 1 to 9 of 9

Thread: How does the BPT position of an object change, as the covering fraction varies?

  1. #1
    Join Date
    Feb 2010
    Posts
    740

    How does the BPT position of an object change, as the covering fraction varies?

    The SDSS fibers - which feed the spectrographs - are 3" in diameter (2" for BOSS). BPT diagrams produced using emission line estimates from these spectra characterize the parts of the galaxies imaged within this keyhole (modulo seeing) ... duh!

    The galaxies may be very much bigger, and the fiber apertures may not be centered on the nuclei (separate issue).

    So an apples-to-apples comparison (of BPT diagram location - SFR-dominated/AGN/LINER/composite/whatever) of many galaxies would require equal - or comparable - fiber covering fractions.

    But SDSS fibers are not custom-made; the covering fraction for galaxy A may be 1%, and for galaxy B 1000% (to turn up the contrast somewhat), so how to 'correct for' the (vastly) different covering fractions?

    IIRC (I'm far too drunk to even care to cite this properly), Kewley et al (2005?) showed that, for 'normal' galaxies, you're more or less OK as long as the covering fraction is <~20% ... but they didn't address position in a BPT diagram (if you don't know what I'm talking about, Google is your Friend!).

    Naively, and excluding EELTs (Hanny's Voorwerp, voorwerpjes, etc), I'd expect that an AGN determined with a covering fraction of 1% (say) could easily become a LINER (say) if said fraction were 100%, and a great many 'composite/transition' objects would end up as 'SFR' (ditto). And if the (1%) fiber just happened to be centered on the equivalent/counterpart of the Tarantula Nebula (or a nuclear starburst), a 100% fiber might have said 'no way! SFR << {your fave elliptical}!' But I have not been able to find any papers on this.

    Anyone?

  2. #2
    Join Date
    Sep 2006
    Posts
    1,826
    Sorry, I don't know of any papers on this myself, but I think that you can answer the question for yourself. Create a model of a galaxy: AGN, bulge, disk, halo, HII regions. Assign a representative spectrum to each component, and decide what fraction of the area of the galaxy is covered by each component. Then start rolling a die to come up with the distance to the galaxy (apparent angular size) and the location of the fiber. Repeat a few thousand times. Note the classification you'd give to the galaxy in each instance, or the position it would have in the BPT diagram.

  3. #3
    Join Date
    Dec 2004
    Posts
    14,782
    For others...

    Abbreviations.com has 22 definitions of "BPT", none of which
    are relevant. It stands for "Baldwin, Phillips & Terlevich''.

    -- Jeff, in Minneapolis

  4. #4
    Join Date
    Aug 2006
    Posts
    321
    There is a paper by Gerssen on on aperture effects where they use integral field spectroscopy for a sample of Halpha emmiters selected from the SDSS:

    http://arxiv.org/abs/1110.6447

    The results are pretty eye opening but with only 24 galaxies and a pretty biased selection, there is plenty more that can and will be done. On that, the SAMI galaxy survey will obtain spatially resolved spectra of around 3000 galaxies within the next few years. Aperture effects will definitely be looked at!

    SAMI = the Sydney university-Australian Astronomical Observatory multi-object integral field spectrograph. It can target 13 galaxies simultaneously. See:

    http://sami-survey.org
    Last edited by matt.o; 2013-Nov-12 at 01:11 PM.

  5. #5
    Join Date
    Feb 2010
    Posts
    740
    Thank you very much matt.o!

    Yes, that's the kind of paper I searched for. By a cosmic quirk of fate, it's one of the many (127, according to ADS) which cite Kewley et al. (2005), so I guess I wasn't as diligent as I should have been in searching the literature.

    Another paper which also cites Kewley et al. (2005) is Iglesias-Páramo et al. (2013):
    Aperture corrections for disk galaxy properties derived from the CALIFA survey. Balmer emission lines in spiral galaxies

    This work investigates the effect of the aperture size on derived galaxy properties for which we have spatially-resolved optical spectra. We focus on some indicators of star formation activity and dust attenuation for spiral galaxies that have been widely used in previous work on galaxy evolution. We investigated 104 spiral galaxies from the CALIFA survey for which 2D spectroscopy with complete spatial coverage is available. From the 3D cubes we derived growth curves of the most conspicuous Balmer emission lines (Hα, Hβ) for circular apertures of different radii centered at the galaxy's nucleus after removing the underlying stellar continuum. We find that the Hα flux (f(Hα)) growth curve follows a well-defined sequence with aperture radius that shows a low dispersion around the median value. From this analysis, we derived aperture corrections for galaxies in different magnitude and redshift intervals. Once stellar absorption is properly accounted for, the f(Hα)/f(Hβ) ratio growth curve shows a smooth decline, pointing toward the absence of differential dust attenuation as a function of radius. Aperture corrections as a function of the radius are provided in the interval [0.3, 2.5]R50. Finally, the Hα equivalent-width (EW(Hα)) growth curve increases with the size of the aperture and shows a very high dispersion for small apertures. This prevents us from using reliable aperture corrections for this quantity. In addition, this result suggests that separating star-forming and quiescent galaxies based on observed EW(Hα) through small apertures will probably result in low EW(Hα) star-forming galaxies begin classified as quiescent.
    @StupendousMan: that's certainly an interesting project (though I'm not sure I have the skills required to do it); however, as the trigger for my question is the GZ Quench Project, I was looking for empirical studies ... if such studies conclude that extrapolating from 'BPT diagram location' of small apertures (relative to total galaxy flux, in the wavelength range of SDSS spectra) to whole galaxies is, um, 'fraught', then naive use of such extrapolations for extremely rare objects (e.g. post-quench galaxies) is surely fraughtfraught.

    @Jeff Root: yes, a BPT diagram - or the kind which plots log ([NII]6584/Hα) against log ([OIII]5007/Hβ) - is what I was referring to. It's a diagnostic tool: an object's position in the diagram says 'here be an AGN/LINER!' or 'star-forming region(s) dude!' or 'bit of both/fence-sitter'. Needless to say, there are traps in deriving each of the x and y values; for example, both the Hα and Hβ emission line flux is supposed to be corrected for stellar absorption (some of the stars sampled by the spectrum have Hα and/or Hβ in absorption; the BPT as a diagnostic needs to 'correct for' this; highly non-trivial).

  6. #6
    Join Date
    Aug 2006
    Posts
    321
    Okay, maybe a 2012 paper by Mike Pracy will be of use to you. The study looks for radial gradients in balmer line absorption using IFU observations of a sample of poststarburst galaxies selected from the SDSS (different mechanisms for the triggering of the initial starburst and the rapid truncation of the star formation predicet differences in the radial gradients of the balmer absorption). Some of them had weak Halpha emission lines but the spatial distribution of the NII/Halpha line strength ratio shows the hard emission is centrally concentrated. There is also an interesting plot on the classification of poststarbursts and the impact of aperture effects, particularly where the radial gradients in balmer absorption are steep. See:

    http://arxiv.org/abs/1111.3371

  7. #7
    Join Date
    Mar 2004
    Posts
    3,161
    Quote Originally Posted by Jean Tate View Post
    \Needless to say, there are traps in deriving each of the x and y values; for example, both the Hα and Hβ emission line flux is supposed to be corrected for stellar absorption (some of the stars sampled by the spectrum have Hα and/or Hβ in absorption; the BPT as a diagnostic needs to 'correct for' this; highly non-trivial).
    Indeed, and one can get very involved in working out the correction. Oddly enough, values from the appendix of this 30-year-old paper continue to be cited.

  8. #8
    Join Date
    Feb 2010
    Posts
    740
    Quote Originally Posted by matt.o View Post
    Okay, maybe a 2012 paper by Mike Pracy will be of use to you. The study looks for radial gradients in balmer line absorption using IFU observations of a sample of poststarburst galaxies selected from the SDSS (different mechanisms for the triggering of the initial starburst and the rapid truncation of the star formation predicet differences in the radial gradients of the balmer absorption). Some of them had weak Halpha emission lines but the spatial distribution of the NII/Halpha line strength ratio shows the hard emission is centrally concentrated. There is also an interesting plot on the classification of poststarbursts and the impact of aperture effects, particularly where the radial gradients in balmer absorption are steep. See:

    http://arxiv.org/abs/1111.3371
    Thanks very much!

    The last column in Figure 3 is a real eye-opener; those gradients are awesome.

    Three, maybe four, of the spectra (Figure 2) seem to have an emission line (or line complex) red-ward of Halpha, do you think that's [SII]? or just some annoying telluric line?

    Anyway, the take-away is that aperture effects cannot be ignored!

    SAMI looks like a very cool project ; do you need any citizen scientists on the team?

  9. #9
    Join Date
    Feb 2010
    Posts
    740
    Quote Originally Posted by ngc3314 View Post
    Indeed, and one can get very involved in working out the correction. Oddly enough, values from the appendix of this 30-year-old paper continue to be cited.
    Maybe because it is a spectacularly good piece of research, and written up well?

    Which leads to this question: is there a relatively simple, straight-forward technique (a.k.a. something you can turn into an algorithm that's codeable - codable? - easily) which be used to derive - correct for - H-Balmer line stellar absorption (principally Halpha and Hbeta) in SDSS spectra? One which works equally well for the hardest of AGNs, the most extreme of SFRs, etc?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •